I realized that my comments on the many suggestions made on the last video could come across as somewhat harsh. This was not my intention, and I intend to encourage discourse and community. I just found it funny how many people had the exact same suggestions. And my joking is often steeped in thick sarcasm, so I can see how it could be misconstrued.
I thought it was very important and interesting to point out the misconceptions present in the community, but maybe next time you could approach it more positively? idk
Glad you posted this. This is the first of your videos I have come across and your reaction to comments certainly put me off at first. Clarification helped a lot.
Is it not backdrivable because the gearteeth would shear, or because there is too much friction, or something else. A cage at the bottom + lubrication would make it even better.
People commenting without looking at other comments may be annoying. But, comments are engagement, right? So, it works out for your video promotion on UA-cam. 😉👍 I have thought UA-camrs should put very silly "oversights" in videos specifically to increase comments/engagement. In your case, it necessitated a follow-up video, too! 🙂
Omg I’m with you, what you don’t see is all the “shit, that didn’t work at all. Time to spend another 2-3 hours modeling and then wait most the day for the damn thing to print.” But it’s so worth it, it’s always nice to come away from so much effort with something to show for it.
The next short one could be a redesign of the gears as the standard spur gears don't have as much surface area as helical gears which should give a bit more strength. Other then that awesome to see you followed up on the original failure! :)
Regular spur gears are actually considered stronger than helical gears. The benefit of helical gears is a higher contact ratio, which is a way of measuring how many teeth make contact at the same time. The difference though is that spur gears make contact along the whole tooth, but helical gears make contact at a point. So in essence, helical gears are better for high speed applications, and they make less noise. For all out high torque applications like this, I think spur gears are the best bet.
PEEK is extremely hard to print, though. I am planning a 3d printer with an awesome, super controlled enclosure which MIGHT be up to the task, but even the pros here on UA-cam barely touch it. Getting a decent benchy to print in PEEK is probably the biggest flex one can pull off in the 3d printing space. It's a bit like being able to play Mozarts Piano Sonata :D
They’d last damn near forever. Those materials are almost about frictionless when in contact with eachother. But, you need special equipment to print with those materials.
This is really cool and inspired me to make one of my own. I was able to make mine with a free wheeling sun gear on the output because I kept the module the same on the input and output. I made it with 50 teeth on the input ring, 20 on the input sun, and 51 on the output ring. This means the input plants were 15 teeth and the output planets were 16 teeth to keep them concentric with the input planets. A small correction to what you stated in the video is that for evenly spaced plants, (sun teeth + ring teeth)/(# planets) must be an integer, not (ring teeth)/(# planets) (you can also have an uneven planet spacing, in which case this does not have to be an integer). This is fortunate because it means that no matter what tooth offset you have between the ring gears, they will both satisfy this equation if one does. Adding some number of teeth t to the input ring means that the output sun will have (input sun - t) teeth. Then we have (s + r)/(# p) = (s-t+r+t)/(# p). t can be any number of teeth, so this would also work for your configuration of 50 on the input and 49 on the output. I found this document to be extremely helpful in learning about gears (page 61 has formulas for planetary gears) qtcgears.com/tools/catalogs/PDF_Q420/Tech.pdf
I realized that the reduction ratio will be less for this version since the output planets have different numbers of teeth, though it is still large. The number of teeth offset per rotation of the carrier is Ri*Po/Pi - Ro. Dividing the output ring teeth by this will give the reduction for just the ring offset. This can be simplified to (Ri+t)*Pi/(t*(Ri-Pi)) where t is the number of teeth difference between the rings. The reduction for the input planetary gear set is still (Ri/Si - 1). So for 50 teeth on the input and 51 on the output, you'd have a 76.5:1 reduction. And for 50 on the input and 49 on the output it would be 73.5.
If you want to maintain ~172:1 and still have 5 planets, but also have an idler sun gear in the center, use the following: Drive Sun - 22 Drive Planet - 18 Fixed Ring - 58 Idler Sun - 31 Output Planet - 24 Output Ring - 79 This will result in 172.36:1 Also, slight correction to your statement at 2:50 - The number of teeth in the ring gear PLUS THE NUMBER OF TEETH ON THE SUN GEAR must be divisible by the number of planets. Since you have 20 teeth in the sun, which is also divisible by 5, it just so happens that if your ring is also divisible by 5, so is the sum. And the problem with your idler sun not fitting at 3:30 is not an inherent property of the mechanism, it's just a result of the tooth counts you have used. Excellent video, I hope to see more on this topic!
If you go up to 6 planets, you can use the follow: Ratio: -171.42857142857204 :1 Planets: 6 Drive Sun: 21 Drive Planet: 15 Fixed Ring: 51 Idler Sun: 24 Output Planet: 18 Output Ring: 60 I'm not sure if this would be stronger than what you already have or not. The teeth on the output side will be a little smaller, but you'll have an extra planet, and the advantage of an idler sun, which might help take a little extra stress maybe.
It turns out it is possible to have an idler sun gear on the top layer if you use the right combination of teeth. If you haven't seen it yet, make sure to watch the video *"Split Ring Compound Planet Epicyclic Gear"* on David Hartkop's channel. I'll add a direct link as a separate reply in case it gets caught in a spam filter.
Really great work! One thought that has sort-of been raised by others is to have an output bearing located on the gearbox output before the lever arm. That should remove most of the load perpendicular to the gearbox and hence the deflection and subsequent damage. Looking at all the BLDC gearboxes I've worked with, they all have a final bearing on the output shaft for this purpose.
I would imagine the start of that failure was 2 of the planet gears getting out of even spacing. Likely that is what stalled the stepper motor. The nature of that mechanism is to forcing the planet gears off axis. The torsion takes them out of parallel. A carrier on both ends would help some, but if you take a look at automobile transmissions you see a single carrier that completely enclose the planet gears. This maintains the parallel of the gears. It may require too much space to be able to be used in your mechanism. You are a very bright young man. Likely you may have already figured it out for yourself.
The addition of the carrier ring is a good solution - but it is a tension ring allowing one planet to move and pull all the others off centre with it. A compression ring would probably be more effective as this would close the gap between it and the outer ring. This introduces a friction element that could be minimized with grease or further reduced by introducing a ball race similar to my 'lazy susan' comment on your cycloidal video. The difference is subtle but would remove the twisting forces on the planet bearings that also failed.
This is kinda what I was thinking after the first video. Adding bearings to the top of the planet gears and bringing the outside in to meet them. Then add another free spinning wheel from the input shaft out to meet it from the inside. Not sure if that would be better than a carrier or your suggestion.
@@dwp6x9e42 bearings on the top of the planet gears running on the outer ring with a compression ring would work. All the lateral loads are cancelled and the teeth become the object of failure testing.
Just a thought...but if you make the top half of the gears longer...and make it so that they only have teeth in the middle section...and a smooth part at the top and bottom of that half...you could place 2 bearings on the 10mm shaft that just roll with the planet gears pushing them outwards and you won't need a planet carrier then.
The best of these use specific planetary mixes from a planetary gear sim that are coherent, a sun gear in the middle of both, and the modulus of both is changed so the planets align, so one planetary modulus is 1 and the other is 1.037 and so the centers of the the planetaries are aligned.
I wrote to makers muse, and I'll write the same to you... Try use POM filament... It's freaking hard to print (wont stick, it'll warp A LOT) but damn... When you manage to print it... that's some GOOD gear material...
Nice video. In the first one you explained that the material strength would be a limiting factor. I would like to see a balanced ratio for the material/motor to see what is the best speed/torque you could get with pla. What would be the best theoretical ratio for steal gears?
Good build... I have to go back and watch what you had going on before because I don't want to say the obvious problem without seeing that you've already been working to tackle that.
so now I know that I want to use planetary screw drives on my exoskeleton elbows, and ATV winch-driven block and tackle cable systems inside the upper arms for heavy lifting with a hub motor belt for regular motion using something similar to this adapted to purpose.
@@LeviJanssen Oh, for sure! You early audio issues are long gone and I'm still convinced you have what it takes for further growth. Even your "failures" are quite interesting! Any colabs moving up on the horizon? It seems those are a valuable way of reaching cross pollination!
@@mattiasfagerlund It seems like every video I make results in a job offer or something crazy like that, but no, no collaborations on the horizon. I struggle to think of myself as being in the caliber of other UA-cam makers.
What if you added a bearing that went around the outside of the carrier you added? Deflection was probably to blame for a lot of the damage. If you could stop all of the gears moving it may have hung in their longer.
I wonder if you have captured on camera the carrier moving as it failed? Did it displace from the central position just before it failed? If so, could you add an additional bearing that guarantees that the new carrier stays located and maybe reduces the risk of the planetary gears twisting ?
Hey man, maybe you won’t see this but what the heck. I’d love to see not only the finished product but how you modeled your pieces, all of your stuff is so cool and sometimes it’s a bit hard to grasp what’s going on right away, I feel if you showed how you made the parts in cad not only would that help people like me grasp it, but inspire design in others.
Levi, for your second ring gear (output), did you keep the same modulus, root circle, and such but literally only reduce teeth count by one? I'm just wondering how you deal with the planetary gears wanting to be "Leaned in" towards the drive shaft of the Stepper motor due to the upper ring gear being a slightly smaller diameter. Maybe I'm over thinking it and the deflection is negligible, but it'd be cool to know what your approach was to remedy this problem. I'm working on building a 3-joint Robotic arm with my Engineering club in College. I've been put in charge of coming up with building a compact gear drive, and boy did your idea hit the jackpot! I created parametrically driven gears in Inventor, but I'm still trying to figure out if there's any special "Tweak" I need to do to either the second half of the Planet gears or Secondary (output) ring gear to accommodate for the change in diameter. Here is how far I got with my design. I also did the carrier plate, but I didn't use a closed stepper motor loop as you did. Looks like you had a NEMA 23 motor. I'm targeting similar for my final robot design. ua-cam.com/video/7ud-kXZYlgg/v-deo.html&ab_channel=TimothySands Cheers!
So can we use this the other way round? To make something spin very fast with a relatively slow but powerful engine. In general with electric motors RPM is not an issue, but still, could be fun to make something spin at insane speed :)
If you don't mind increasing the height, you could put a third level of planet teeth matching the sun teeth, and extend the sun gear with teeth at the first and third levels
Yes, this is what I was thinking too. You don't necessarily need to make it taller either, if you use half size on the top and bottom. Some cycloidal projects do this.
@@hughsanderson1834 Doing a split primary to encapsulate the secondary would also ideally have a ring around the upper primary, which would require the output ring to have a larger flange; though, this larger flange could also allow the bolted-on arm to encapsulate a thin bearing for the upper primary, so maybe that is a worthwhile investment. This design would still need a carrier at each end, which would need to be properly located and not free-floating like the current design has; this design flaw, the free-floating carrier, is what allowed one planet that skewed to pull every planet out of alignment. Any epicyclic that uses floating planets is just asking for failure, and a floating carrier isn't much better. Though, I'd also be curious about doing the inverse of this, central primary with a split secondary. The lower ring would be encapsulated by the primary's central ring's connection to the motor's flange, and a bearing here would be nice as well. This would also allow for the same slimmer profile of the output, and puts the output on top. This might even provide a nicer final design, as it can all be encapsulated by a box with the same square dimension as the motor, everything enclosed, with the output on top.
If you really wanna smear those teeth, maybe what I suggested on the prior video would work? (Instead of a carrier, have a bearing on the center hub, and bearing on the top of each red gear. That way the failure mode will be to really flatten the outer and inner red gears against each other!)
Instead of putting a gear in the top sun plane you could just put a spacer/bushing/"toothless gear" or whatever you may call it for the teeth of the planetary gears to ride on instead of the overly complicated carrier.
The tips of the gear teeth aren't designed to take that kind of loading. Not sure why you describe the carrier as "complicated" either, because it's pretty simple...
@@LeviJanssen BTW, I over exaggerated the ratio. The one I saw only had a 70.4:1 ratio and is limited by mechanical constraints, where yours is more flexible since it doesn't use an idler gear.
I'm all but sure I'm missing something, but why does the hypothetical upper separator ring need teeth at all? It's there only to provide outward force, not torque.
Your failure was asymmetrical, suggesting that, even if you used all-metal parts and nothing failed, you're putting asymmetrical pressure on your gears. Is that ideal?
Could a second sun gear, set of planets & SPLIT RING be added on the other side of the driven ring? You'd be supporting the driving planets from either side, essentially eliminating the bending of the planet gears, which seems to be the main source of failure at the moment.
Could you apply the weight, but use the gear box reverse to produce high rpm to a motor to produce power? “Gravity Battery”. Similar to how a coo coo clock works. The weight falls slow and the center fast.
Nice...so discounting the material strength...the actual vs the real world, it get's approximately 10% of calculated mathematical torque spec multiplication before failure Are the inner gears fixed together...which means due to the offset in manufacture they probably couldn't be manufactured with traditional methods unless the gears themselves had a space between the top/bottom portion with a space between them. Still so very awesome and very spectacular failure
What about adding grease, so power lost on friction can go to the output shaft? From video of Let's print "3d printed worm gear" i got that grease makes huge difference.
Thanks for the follow up. On the topic of the idler gear, that got me going down a rabbit hole to learn... well, a bit more then the nothing I knew before. I think your explanation is poorly worded: you say that the ring gear tooth count needs to be divisible by number of planets for meshing- but from reading it's that ring+sun need to be divisible by number of planets (and possibly evenly divisible to get equal spacing of planets?)- (for anyone else looking to learn: ua-cam.com/video/71dn-eVdSmc/v-deo.html has most of a nice 30 second explanation in the middle of a 40 minute video- basically tooth counts of the various gears are directly correlated to diameters, and the diameter of ring needs to be equal to diameter of 2 planets + 1 sun). That accounted for, your argument does still make sense: 49+20 isn't divisible by 5 either, so, yeah, idler was never going to work for the current tooth counts / if your goal is to keep the off-by-one harmonic geometry. The 'ring gear must be divisible by planet count' argument had me confused / I first heard it as a more general "compound planetarys can't have idlers" which led me down trying to figure out why some compound planetary can have idlers and this one couldn't- ex: I've got a copy of Gear Down for What's 1379:1 single stage reducer ( www.thingiverse.com/thing:2220506 ) that does have an output idler (and 59T ring w/ 3 planets). From looking at the specs it seems like there's a lot of fiddling with the module and tooth counts to get the planets of both planetary gear sets to line up, and that probably does limit what can be done as far as ring gear tooth counts, which I _think_ is what you were arguing in other comments? I'm already out of my depth, so it's a bit hard to follow. And, now that I understand this a little bit more, I've got a question: for the 49T output ring, did you use a slightly larger module in order to get it to the same diameter as the 50T ring? Or maybe a different module on the planets? Or not accounted for, and they just bend in slightly? Anyway, thanks for the video/follow up/inspiration-to-read-more-on-gearsets; and hope you're having fun with the Austin internship!
"I set out today to break something, and I broke something." Doesn't it just feel great when your day goes exactly as planned? xD Still, I think this is a pretty incredible showing for such a small gearbox. The materials might not be able to handle very large forces, but I can see some applications where you'd want to boost the torque of, like, a tiny motor. Nothing industrial, but maybe a tiny RC drone for work in crawlspaces or ducts or something.
@@LeviJanssen I've seen some UA-camrs making attempts to build their own. The difficult part seems to be the power supply. Edit: Ben Krasnow got his PSU from BAX EDM ua-cam.com/channels/y4kgsAYxcraee8w5SfqXPA.html And the Idahoan Show successfully made a die sink type with a far simpler PSU. ua-cam.com/video/6Jt08F1HOiU/v-deo.html
It might be better to partner with someone or a company that already has the equipment. When the design is final maybe this old Tony is looking for a project: ua-cam.com/video/Q-XOM4E4RZQ/v-deo.html. If you want to put them into production a Kickstarter to have a larger quantity produced may work. It may also be possible to get off the shelf planet/sun gears that key to a shaft and just print the ring gear and other components.
Hello, I am Brazilian with physical disabilities, I am adapting my tractor. with these gears will it be possible to lift a 6m bar weighing 40kg from a position of 180° to 90°?
The reason why people suggest things that have already been suggested is that UA-cam comment are an opaque, unusable and unreadable mess. They simply could not see them. There is no way to get a sense of what has already been said unless you spend a ridiculous amount of time wading through the comments, which nobody will do before commenting. And if they had to, they would simply not bother commenting in the first place.
Adding a second sun gear is possible though, I did a project to test it : www.thingiverse.com/thing:4889174 The second sun need to have a different number of teeth (but it doesn't matter as it is idle) and the geometric constraints are being solved by backlash gaps and plastic flexibility.
@@LeviJanssen dude wtf? Its not about being right. Everyone is here to learn, I hope you are too. Why so rough? Maybe I understand you wrong, but also your opening remarks on people that follow you, engage with you and your content and try to be helpful sounded borderline insulting and at least condescending. Wjy would you be like that? Also this I'm writing not to argue with you, but because I hope it will be helpful feedback. I do like your content and I do care :3
@@Mosenhosen That was not my intention at all. My joking tends to involve a lot of dry sarcasm. Check out the pinned comment, and I appreciate your feedback.
I am dumb as a rock .So forgive me if what I say is shear stupidity . Why would a middle support have to have teeth .Why not a smooth wheel that the teeth just ride on top of to keep everything lined up?
You can actually add a 2nd sun gear by changing the size of the planets on the 2nd stage and size of sun and ring accordingly, check out this video: ua-cam.com/video/2dG_F9rR-jM/v-deo.html he doesn't explain a lot but its possible, here is another video with the math even if its not ituitive at all : ua-cam.com/video/5a1w9daIybc/v-deo.html
Like your videos, but the edits in the beginning were a bit nauseating for me. Try using that zoom in/out a bit less. I know you want to emphasize what you were saying, but it was just way too much too quickly. Less is more, and fewer cuts are better.
You are still a small channel. You need to get used to people just coming to your comment section and making their comment, having no regard to what was already written there. There is a lot more people interested in sharing their thoughts, then there is people interested in the toxicity/trolling/stupidity of the regular you tube comment section, or reading through hours of comments. Yes maybe your channel is not so bad, but it is a reality on most channels. Not only that but we are told it is beneficial to your channel having more comments, even if they are basically the same dumb suggestion x100. So just embrace that, more comments means more exposure for you (so we are told). Don't let the comment section cause any negativity to you. *me goes back to watching vids and leaving no comments on most of them* nb: on yeah you pinned your own comment there explaining that thing. well I didn't even notice it until after I posted. But I did you no harm, look +1 comment on your comment section.
It's neat that you make these videos, but isn't this whole series basically just a long-winded wrong answer on an ME exam on gear-sets? (49 is not divisible by 5, quelle surprise!!) I came across this vid after watching a similar one posted 5 years ago that even provided a calc spreadsheet to get it right.
I realized that my comments on the many suggestions made on the last video could come across as somewhat harsh. This was not my intention, and I intend to encourage discourse and community. I just found it funny how many people had the exact same suggestions. And my joking is often steeped in thick sarcasm, so I can see how it could be misconstrued.
You can maybe scale the design a little bit and add a bearing in the middle of the planet carrier thing?
I thought it was very important and interesting to point out the misconceptions present in the community, but maybe next time you could approach it more positively? idk
Glad you posted this. This is the first of your videos I have come across and your reaction to comments certainly put me off at first. Clarification helped a lot.
Is it not backdrivable because the gearteeth would shear, or because there is too much friction, or something else. A cage at the bottom + lubrication would make it even better.
People commenting without looking at other comments may be annoying.
But, comments are engagement, right? So, it works out for your video promotion on UA-cam. 😉👍
I have thought UA-camrs should put very silly "oversights" in videos specifically to increase comments/engagement. In your case, it necessitated a follow-up video, too! 🙂
The engineering process is a lot cooler when it’s someone else going through it
Can confirm
Omg I’m with you, what you don’t see is all the “shit, that didn’t work at all. Time to spend another 2-3 hours modeling and then wait most the day for the damn thing to print.” But it’s so worth it, it’s always nice to come away from so much effort with something to show for it.
The next short one could be a redesign of the gears as the standard spur gears don't have as much surface area as helical gears which should give a bit more strength. Other then that awesome to see you followed up on the original failure! :)
Косозубые шестерни, и увеличить размеры устройства - ширина и диаметр, для снижения нагрузки в каждой точке передачи..
Regular spur gears are actually considered stronger than helical gears. The benefit of helical gears is a higher contact ratio, which is a way of measuring how many teeth make contact at the same time. The difference though is that spur gears make contact along the whole tooth, but helical gears make contact at a point. So in essence, helical gears are better for high speed applications, and they make less noise. For all out high torque applications like this, I think spur gears are the best bet.
You need a cage at the bottom of that planetary set as well. Might last better.
My favorite line: "I wanted to see some gears get turned into wheels..." I'm really enjoying seeing your design/test/improvement process, keep it up!
"...some gears get turned into whales..." and yes I had to go back a couple of times to figure out what he said
I'm really curious to see how these gears would hold up if they were made of a stronger 3d printed material (e.g. Nylon, PEEK, Ultem, etc.)
PEEK is extremely hard to print, though. I am planning a 3d printer with an awesome, super controlled enclosure which MIGHT be up to the task, but even the pros here on UA-cam barely touch it. Getting a decent benchy to print in PEEK is probably the biggest flex one can pull off in the 3d printing space. It's a bit like being able to play Mozarts Piano Sonata :D
They’d last damn near forever. Those materials are almost about frictionless when in contact with eachother. But, you need special equipment to print with those materials.
Herringbone gears tend to make 3d printed gear teeth a lot stronger without significant drawbacks...
This is really cool and inspired me to make one of my own. I was able to make mine with a free wheeling sun gear on the output because I kept the module the same on the input and output. I made it with 50 teeth on the input ring, 20 on the input sun, and 51 on the output ring. This means the input plants were 15 teeth and the output planets were 16 teeth to keep them concentric with the input planets. A small correction to what you stated in the video is that for evenly spaced plants, (sun teeth + ring teeth)/(# planets) must be an integer, not (ring teeth)/(# planets) (you can also have an uneven planet spacing, in which case this does not have to be an integer). This is fortunate because it means that no matter what tooth offset you have between the ring gears, they will both satisfy this equation if one does. Adding some number of teeth t to the input ring means that the output sun will have (input sun - t) teeth. Then we have (s + r)/(# p) = (s-t+r+t)/(# p). t can be any number of teeth, so this would also work for your configuration of 50 on the input and 49 on the output. I found this document to be extremely helpful in learning about gears (page 61 has formulas for planetary gears) qtcgears.com/tools/catalogs/PDF_Q420/Tech.pdf
I realized that the reduction ratio will be less for this version since the output planets have different numbers of teeth, though it is still large. The number of teeth offset per rotation of the carrier is Ri*Po/Pi - Ro. Dividing the output ring teeth by this will give the reduction for just the ring offset. This can be simplified to (Ri+t)*Pi/(t*(Ri-Pi)) where t is the number of teeth difference between the rings. The reduction for the input planetary gear set is still (Ri/Si - 1). So for 50 teeth on the input and 51 on the output, you'd have a 76.5:1 reduction. And for 50 on the input and 49 on the output it would be 73.5.
If you want to maintain ~172:1 and still have 5 planets, but also have an idler sun gear in the center, use the following:
Drive Sun - 22
Drive Planet - 18
Fixed Ring - 58
Idler Sun - 31
Output Planet - 24
Output Ring - 79
This will result in 172.36:1
Also, slight correction to your statement at 2:50 - The number of teeth in the ring gear PLUS THE NUMBER OF TEETH ON THE SUN GEAR must be divisible by the number of planets. Since you have 20 teeth in the sun, which is also divisible by 5, it just so happens that if your ring is also divisible by 5, so is the sum. And the problem with your idler sun not fitting at 3:30 is not an inherent property of the mechanism, it's just a result of the tooth counts you have used.
Excellent video, I hope to see more on this topic!
If you go up to 6 planets, you can use the follow:
Ratio: -171.42857142857204 :1
Planets: 6
Drive Sun: 21
Drive Planet: 15
Fixed Ring: 51
Idler Sun: 24
Output Planet: 18
Output Ring: 60
I'm not sure if this would be stronger than what you already have or not. The teeth on the output side will be a little smaller, but you'll have an extra planet, and the advantage of an idler sun, which might help take a little extra stress maybe.
I can't get enough 3d printed gearboxes.
Hello, my long lost brother
@@Olegach21 Hello you must be the guy on the other side of the mirror.
@@madeintexas3d442 get a room you two... 🙄😂
It turns out it is possible to have an idler sun gear on the top layer if you use the right combination of teeth.
If you haven't seen it yet, make sure to watch the video *"Split Ring Compound Planet Epicyclic Gear"* on David Hartkop's channel. I'll add a direct link as a separate reply in case it gets caught in a spam filter.
Here's a link to David's video:
ua-cam.com/video/-VtbSvVxaFA/v-deo.html
in the end, this was a test of plastic, not gears. We learned not to use plastic for a weight-bearing Split-Ring Compound Planetary
Really great work! One thought that has sort-of been raised by others is to have an output bearing located on the gearbox output before the lever arm. That should remove most of the load perpendicular to the gearbox and hence the deflection and subsequent damage. Looking at all the BLDC gearboxes I've worked with, they all have a final bearing on the output shaft for this purpose.
I would imagine the start of that failure was 2 of the planet gears getting out of even spacing. Likely that is what stalled the stepper motor. The nature of that mechanism is to forcing the planet gears off axis. The torsion takes them out of parallel. A carrier on both ends would help some, but if you take a look at automobile transmissions you see a single carrier that completely enclose the planet gears. This maintains the parallel of the gears. It may require too much space to be able to be used in your mechanism. You are a very bright young man. Likely you may have already figured it out for yourself.
Looks good my only recommendations are put do the same on the other side and support them with bearings.
The addition of the carrier ring is a good solution - but it is a tension ring allowing one planet to move and pull all the others off centre with it. A compression ring would probably be more effective as this would close the gap between it and the outer ring. This introduces a friction element that could be minimized with grease or further reduced by introducing a ball race similar to my 'lazy susan' comment on your cycloidal video. The difference is subtle but would remove the twisting forces on the planet bearings that also failed.
This is kinda what I was thinking after the first video. Adding bearings to the top of the planet gears and bringing the outside in to meet them. Then add another free spinning wheel from the input shaft out to meet it from the inside. Not sure if that would be better than a carrier or your suggestion.
@@dwp6x9e42 bearings on the top of the planet gears running on the outer ring with a compression ring would work. All the lateral loads are cancelled and the teeth become the object of failure testing.
Just a thought...but if you make the top half of the gears longer...and make it so that they only have teeth in the middle section...and a smooth part at the top and bottom of that half...you could place 2 bearings on the 10mm shaft that just roll with the planet gears pushing them outwards and you won't need a planet carrier then.
The best of these use specific planetary mixes from a planetary gear sim that are coherent, a sun gear in the middle of both, and the modulus of both is changed so the planets align, so one planetary modulus is 1 and the other is 1.037 and so the centers of the the planetaries are aligned.
I wrote to makers muse, and I'll write the same to you...
Try use POM filament... It's freaking hard to print (wont stick, it'll warp A LOT) but damn... When you manage to print it... that's some GOOD gear material...
Very nice! It would be great to see this backdriven (assuming the ratio is reduced to make that possible)
It's a shame it's not back drivable. It would be nice to have a simple gravity generator like that.
using a regular planetary set and some pulley shenanigans would be easiest for an application like this I think.
Nice video. In the first one you explained that the material strength would be a limiting factor. I would like to see a balanced ratio for the material/motor to see what is the best speed/torque you could get with pla. What would be the best theoretical ratio for steal gears?
Audio is much clearer than past videos, and your storytelling continues to be great
Good build... I have to go back and watch what you had going on before because I don't want to say the obvious problem without seeing that you've already been working to tackle that.
so now I know that I want to use planetary screw drives on my exoskeleton elbows, and ATV winch-driven block and tackle cable systems inside the upper arms for heavy lifting with a hub motor belt for regular motion using something similar to this adapted to purpose.
Almost 20k subscribers! Well done!
Thanks! Glad to see you’ve stuck around
@@LeviJanssen Oh, for sure! You early audio issues are long gone and I'm still convinced you have what it takes for further growth. Even your "failures" are quite interesting! Any colabs moving up on the horizon? It seems those are a valuable way of reaching cross pollination!
@@mattiasfagerlund It seems like every video I make results in a job offer or something crazy like that, but no, no collaborations on the horizon. I struggle to think of myself as being in the caliber of other UA-cam makers.
Motor stopped working because the planets meshed when there was the crunch sound. It needs a bottom carrier as well!
I love seeing your projects. Separately, it’s also fun to see your storytelling and video production improve with each post.
Thank you for posting this, I learned something about planetary gears from your very clear explanation, thank you.
What if you added a bearing that went around the outside of the carrier you added? Deflection was probably to blame for a lot of the damage. If you could stop all of the gears moving it may have hung in their longer.
I wonder if you have captured on camera the carrier moving as it failed? Did it displace from the central position just before it failed? If so, could you add an additional bearing that guarantees that the new carrier stays located and maybe reduces the risk of the planetary gears twisting ?
youre going to go far in life dude. god bless you.
Hey man, maybe you won’t see this but what the heck. I’d love to see not only the finished product but how you modeled your pieces, all of your stuff is so cool and sometimes it’s a bit hard to grasp what’s going on right away, I feel if you showed how you made the parts in cad not only would that help people like me grasp it, but inspire design in others.
Levi, for your second ring gear (output), did you keep the same modulus, root circle, and such but literally only reduce teeth count by one? I'm just wondering how you deal with the planetary gears wanting to be "Leaned in" towards the drive shaft of the Stepper motor due to the upper ring gear being a slightly smaller diameter. Maybe I'm over thinking it and the deflection is negligible, but it'd be cool to know what your approach was to remedy this problem.
I'm working on building a 3-joint Robotic arm with my Engineering club in College. I've been put in charge of coming up with building a compact gear drive, and boy did your idea hit the jackpot! I created parametrically driven gears in Inventor, but I'm still trying to figure out if there's any special "Tweak" I need to do to either the second half of the Planet gears or Secondary (output) ring gear to accommodate for the change in diameter.
Here is how far I got with my design. I also did the carrier plate, but I didn't use a closed stepper motor loop as you did. Looks like you had a NEMA 23 motor. I'm targeting similar for my final robot design.
ua-cam.com/video/7ud-kXZYlgg/v-deo.html&ab_channel=TimothySands
Cheers!
you should checkout recent Maker's Muse video where he is testing gears out of different filament for strength.
Not necessary it was common knowledge that harder more Brittle materials are better for gears
@@shorb2289 why brittle material is better for gear?
So can we use this the other way round? To make something spin very fast with a relatively slow but powerful engine. In general with electric motors RPM is not an issue, but still, could be fun to make something spin at insane speed :)
If you don't mind increasing the height, you could put a third level of planet teeth matching the sun teeth, and extend the sun gear with teeth at the first and third levels
Yes, this is what I was thinking too. You don't necessarily need to make it taller either, if you use half size on the top and bottom. Some cycloidal projects do this.
@@hughsanderson1834 Doing a split primary to encapsulate the secondary would also ideally have a ring around the upper primary, which would require the output ring to have a larger flange; though, this larger flange could also allow the bolted-on arm to encapsulate a thin bearing for the upper primary, so maybe that is a worthwhile investment. This design would still need a carrier at each end, which would need to be properly located and not free-floating like the current design has; this design flaw, the free-floating carrier, is what allowed one planet that skewed to pull every planet out of alignment. Any epicyclic that uses floating planets is just asking for failure, and a floating carrier isn't much better.
Though, I'd also be curious about doing the inverse of this, central primary with a split secondary. The lower ring would be encapsulated by the primary's central ring's connection to the motor's flange, and a bearing here would be nice as well. This would also allow for the same slimmer profile of the output, and puts the output on top. This might even provide a nicer final design, as it can all be encapsulated by a box with the same square dimension as the motor, everything enclosed, with the output on top.
If you really wanna smear those teeth, maybe what I suggested on the prior video would work? (Instead of a carrier, have a bearing on the center hub, and bearing on the top of each red gear. That way the failure mode will be to really flatten the outer and inner red gears against each other!)
Instead of putting a gear in the top sun plane you could just put a spacer/bushing/"toothless gear" or whatever you may call it for the teeth of the planetary gears to ride on instead of the overly complicated carrier.
The tips of the gear teeth aren't designed to take that kind of loading. Not sure why you describe the carrier as "complicated" either, because it's pretty simple...
I love this mechanism.
Apparently, there are split-ring compound designs used in rearview mirrors that have ratios of 528:1 that are as compact as yours.
Yeah, rearview mirror actuators are the only widespread application I know of that uses this mechanism.
@@LeviJanssen BTW, I over exaggerated the ratio. The one I saw only had a 70.4:1 ratio and is limited by mechanical constraints, where yours is more flexible since it doesn't use an idler gear.
That's really cool man.
You should try making an engine connecting rod that's a compliant mechanism.
Congrats on 20k subs!
I'm all but sure I'm missing something, but why does the hypothetical upper separator ring need teeth at all? It's there only to provide outward force, not torque.
Great video. Is it possible to use helical gears with this?
Your failure was asymmetrical, suggesting that, even if you used all-metal parts and nothing failed, you're putting asymmetrical pressure on your gears. Is that ideal?
Could a second sun gear, set of planets & SPLIT RING be added on the other side of the driven ring? You'd be supporting the driving planets from either side, essentially eliminating the bending of the planet gears, which seems to be the main source of failure at the moment.
TIL PSUs can scream... and I was NOT prepared to hear those screams of agony this loud on something like that. XD
Could you apply the weight, but use the gear box reverse to produce high rpm to a motor to produce power? “Gravity Battery”. Similar to how a coo coo clock works. The weight falls slow and the center fast.
I would love to see this made in metal to see how much it could stand
Would making this with helical gears increase the stall torque?
Косозубые шестерни, и увеличить размеры устройства - ширина и диаметр, для снижения нагрузки в каждой точке передачи..
@@arseniyseleznovveroboj1571 Я понимаю. Это имеет смысл, я бы хотел увидеть более тяжелую и прочную версию этого
You know, somehow im trally surprised youve never made somesort of automatic transmission
Nice...so discounting the material strength...the actual vs the real world, it get's approximately 10% of calculated mathematical torque spec multiplication before failure
Are the inner gears fixed together...which means due to the offset in manufacture they probably couldn't be manufactured with traditional methods unless the gears themselves had a space between the top/bottom portion with a space between them.
Still so very awesome and very spectacular failure
What about adding grease, so power lost on friction can go to the output shaft? From video of Let's print "3d printed worm gear" i got that grease makes huge difference.
Thanks for the follow up. On the topic of the idler gear, that got me going down a rabbit hole to learn... well, a bit more then the nothing I knew before. I think your explanation is poorly worded: you say that the ring gear tooth count needs to be divisible by number of planets for meshing- but from reading it's that ring+sun need to be divisible by number of planets (and possibly evenly divisible to get equal spacing of planets?)- (for anyone else looking to learn: ua-cam.com/video/71dn-eVdSmc/v-deo.html has most of a nice 30 second explanation in the middle of a 40 minute video- basically tooth counts of the various gears are directly correlated to diameters, and the diameter of ring needs to be equal to diameter of 2 planets + 1 sun). That accounted for, your argument does still make sense: 49+20 isn't divisible by 5 either, so, yeah, idler was never going to work for the current tooth counts / if your goal is to keep the off-by-one harmonic geometry.
The 'ring gear must be divisible by planet count' argument had me confused / I first heard it as a more general "compound planetarys can't have idlers" which led me down trying to figure out why some compound planetary can have idlers and this one couldn't- ex: I've got a copy of Gear Down for What's 1379:1 single stage reducer ( www.thingiverse.com/thing:2220506 ) that does have an output idler (and 59T ring w/ 3 planets). From looking at the specs it seems like there's a lot of fiddling with the module and tooth counts to get the planets of both planetary gear sets to line up, and that probably does limit what can be done as far as ring gear tooth counts, which I _think_ is what you were arguing in other comments? I'm already out of my depth, so it's a bit hard to follow.
And, now that I understand this a little bit more, I've got a question: for the 49T output ring, did you use a slightly larger module in order to get it to the same diameter as the 50T ring? Or maybe a different module on the planets? Or not accounted for, and they just bend in slightly?
Anyway, thanks for the video/follow up/inspiration-to-read-more-on-gearsets; and hope you're having fun with the Austin internship!
i thought you were doing this videos in the netherlands, but thats obviously not the case. So where do you record these videos?
Would fitting a bearing on the new "carrier" and the metal shaft help distribute more load through it than just the gears teeth?
This seems like a good idea.
Would this work with herringbone teeth I think that would make it hold more weight. Spreading the load over more contact area
"I set out today to break something, and I broke something." Doesn't it just feel great when your day goes exactly as planned? xD
Still, I think this is a pretty incredible showing for such a small gearbox. The materials might not be able to handle very large forces, but I can see some applications where you'd want to boost the torque of, like, a tiny motor. Nothing industrial, but maybe a tiny RC drone for work in crawlspaces or ducts or something.
Thank you
I know it's quite an investment, but since you're designing your own gears, you might benefit from EDM to cut custom gears out of steel.
I assume most would benefit from wire EDM, but most don’t have a million dollars laying around
@@LeviJanssen I've seen some UA-camrs making attempts to build their own. The difficult part seems to be the power supply.
Edit: Ben Krasnow got his PSU from BAX EDM ua-cam.com/channels/y4kgsAYxcraee8w5SfqXPA.html
And the Idahoan Show successfully made a die sink type with a far simpler PSU. ua-cam.com/video/6Jt08F1HOiU/v-deo.html
It might be better to partner with someone or a company that already has the equipment. When the design is final maybe this old Tony is looking for a project: ua-cam.com/video/Q-XOM4E4RZQ/v-deo.html. If you want to put them into production a Kickstarter to have a larger quantity produced may work. It may also be possible to get off the shelf planet/sun gears that key to a shaft and just print the ring gear and other components.
Hello, I am Brazilian with physical disabilities, I am adapting my tractor. with these gears will it be possible to lift a 6m bar weighing 40kg from a position of 180° to 90°?
could you try helical gears they should be stronger and not as loud at high speeds
Hey, sorry for botherong you but your first idea with the coil gun,(thicker coil at the end) would it work?
3rd part, please. ;)
9:30 "I wanted to see some gears get turned into whales" lol
The reason why people suggest things that have already been suggested is that UA-cam comment are an opaque, unusable and unreadable mess.
They simply could not see them. There is no way to get a sense of what has already been said unless you spend a ridiculous amount of time wading through the comments, which nobody will do before commenting. And if they had to, they would simply not bother commenting in the first place.
same two-side balance fix for the outer gearing
try v-teeth instead
Are you stepping it too fast so the current (and thus torque) can't build up to t's full capacity?
Nice video !
no print files?
Hello, I'm Brazilian, disabled, I'm trying to adapt my tractor. Do you sell this project so I can print it?
Very interesting
Why does the camera jump around sometimes when you say just one word?
This structure, How many GEAR BACKLASH ？
What was the reduction on this?
i'm a bit uncomfortable with "i want to see failure." Would it be better to say "I want to find, and fix, the flaw in my design."?
Steel Gears are next..... and this gearbox will take over the robot world.... ;-)
Carrier should have better output tourque
Adding a second sun gear is possible though, I did a project to test it : www.thingiverse.com/thing:4889174
The second sun need to have a different number of teeth (but it doesn't matter as it is idle) and the geometric constraints are being solved by backlash gaps and plastic flexibility.
That's fascinating.
As you say, the teeth don’t mesh. I suppose anything can fit if it bends enough, but it’s not at all desirable.
@@LeviJanssen dude wtf? Its not about being right. Everyone is here to learn, I hope you are too. Why so rough?
Maybe I understand you wrong, but also your opening remarks on people that follow you, engage with you and your content and try to be helpful sounded borderline insulting and at least condescending. Wjy would you be like that?
Also this I'm writing not to argue with you, but because I hope it will be helpful feedback. I do like your content and I do care :3
@@Mosenhosen That was not my intention at all. My joking tends to involve a lot of dry sarcasm. Check out the pinned comment, and I appreciate your feedback.
@@LeviJanssen 👍 I understood that it can be funny for some, maybe even the majority of your viewers, still it was a joke on the expense of others.
Heyy... Did you finished the coil gun project that you did videos before..... I was waiting for that for a long time
What type of filament are you using?
Normal PLA
@@LeviJanssen ok, thanks!
Will you use it for something? =)
I am dumb as a rock .So forgive me if what I say is shear stupidity . Why would a middle support have to have teeth .Why not a smooth wheel that the teeth just ride on top of to keep everything lined up?
Great explaintion why the freewheel sun doesn't work
You can actually add a 2nd sun gear by changing the size of the planets on the 2nd stage and size of sun and ring accordingly, check out this video:
ua-cam.com/video/2dG_F9rR-jM/v-deo.html
he doesn't explain a lot but its possible, here is another video with the math even if its not ituitive at all :
ua-cam.com/video/5a1w9daIybc/v-deo.html
Changing these relationships would fundamentally change the mechanism and reduce the reduction ratio
@@LeviJanssen ua-cam.com/video/-VtbSvVxaFA/v-deo.html
maybe with a configuration like this? planet teeth don't change but diameter do
Like your videos, but the edits in the beginning were a bit nauseating for me. Try using that zoom in/out a bit less. I know you want to emphasize what you were saying, but it was just way too much too quickly. Less is more, and fewer cuts are better.
You are still a small channel. You need to get used to people just coming to your comment section and making their comment, having no regard to what was already written there.
There is a lot more people interested in sharing their thoughts, then there is people interested in the toxicity/trolling/stupidity of the regular you tube comment section, or reading through hours of comments. Yes maybe your channel is not so bad, but it is a reality on most channels. Not only that but we are told it is beneficial to your channel having more comments, even if they are basically the same dumb suggestion x100. So just embrace that, more comments means more exposure for you (so we are told). Don't let the comment section cause any negativity to you. *me goes back to watching vids and leaving no comments on most of them*
nb: on yeah you pinned your own comment there explaining that thing. well I didn't even notice it until after I posted. But I did you no harm, look +1 comment on your comment section.
put some grease on it
2:18 Pretty naive of you to think people read the comments
the only failure i see are your people skills
It's neat that you make these videos, but isn't this whole series basically just a long-winded wrong answer on an ME exam on gear-sets? (49 is not divisible by 5, quelle surprise!!) I came across this vid after watching a similar one posted 5 years ago that even provided a calc spreadsheet to get it right.
your entire videos are out of focus it's triggering my sort of OCD
U need a mill !! or SLS Bruh!! that's not u'r foot? U'r teeth look different? find a app that sets u'r angles..was that u'r foot?
Feel free to try "your" next time instead of "u'r", which is just plain weird
@@clonkex aint it!🌮👀